SPINE Volume 39, Number 12, pp 963-970 ©2014, Lippincott Williams & Wilkins
HEALTH SERVICES RESEARCH
Translation, Cross-cultural Adaptation, and Validation of a Chinese Version of the Japanese Orthopaedic Association Cervical Myelopathy Evaluation Questionnaire Andy Chien, PhD,*+ Dar-Ming Lai, PhD,* Chih-Hsiu Cheng, PhD,§ Shwu-Fen Wang, PhD,t Wei-Li Hsu, PhD,t and Jaw-Lin Wang, PhD*
Study Design. Translation and psychometric testing of a questionnaire. Objective. To adapt the Japanese Orthopaedic Association Cervical Myelopathy Evaluation Questionnaire QOACMEQ) crossculturally to Chinese language and to investigate the psychometric properties ofthe translated Chinese lOACMEQ. Summary of Bacitground Data. The recently developed JOACMEQ is a self-reported questionnaire used to assess the severity of the cervical spondylotic myelopathy. Despite its acceptance in the literature, the JOACMEQ has not been translated and validated for Chinese-speaking patients. M e t h o d s . Translation and cross-cultural adaptation of the original JOACMEQ was performed according to published guidelines. A total of 105 patients with a predominant complain of neck pain that included but not limited to patients with a clinically confirmed cervical spondylotic myelopathy were included. Psychometric testing of the responses included construct validity (factor analysis), internal consistency (Cronbach a), test-retest reproducibüity (intraclass correlation coefficient), and concurrent validity (Spearman p). The discriminatory power of differentiating those with myelopathy from
From the 'Institute of Biomédical Engineering, National Taiwan University, Taipei, Taiwan; tSchool and Graduate Institute of Physical Therapy, College of Medicine, National Taiwan University, Taipei, Taiwan; ^Department of Surgery, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan; and §Department of Physical Therapy and Graduate Institute of Rehabilitation Science, Chang Gung University, Taoyuan, Taiwan. Acknowledgment date: January 2, 2014. Revision date: February 19, 2014. Acceptance date: February 24, 2014. The manuscript submitted does not contain information about medical device(s)/drug(s). The National Science Council, Taiwan (NSC 101-2221-F-002-060-MY3) grant funds were received to support this work. No relevant financial activities outside the submitted work. Address correspondence and reprint requests to |aw-Lin Wang, PhD, Department of Mechanical Engineering, College of Engineering, National Taiwan University, 602 |en-Su Hall, 1 Section 4, Roosevelt Rd, Taipei 10617, Taiwan, ROC; E-mail: [email protected] DOI: 10.1097/BRS.0000000000000313 Spine
those without was assessed by the area under the receiver operating characteristic curve and with the optimal cutoff points determined. Results. Construct validity testing revealed a comparable factorial structure of the Chinese JOACMEQ with the original JOACMEQ. The Cronbach a of the Chinese JOACMEQ was 0.88, indicating good internal consistency. Good to excellent test-retest reliability was demonstrated for all 5 domain scores (intraclass correlation coefficient range = 0.79-0.90). A significant strong association between the Chinese Neck Disability Index and the quality of life domain of the Chinese JOACMEQ (Spearman p = - 0 . 7 6 , P < 0.01 ) was identified. The upper and lower extremity function domain scores were found to be significant, with the area under the receiver operating characteristic curve found to be 0.75 {P = 0.04) and 0.82 (P = 0.02), respectively. Conclusion. The results indicated that the Chinese JOACMEQ successfully retained the psychometric properties of the original JOACMEQ and support the usefulness of the Chinese JOACMEQ as an appropriate supplementary diagnostic and outcome measure for Chinese patients suspected of cervical spondylotic myelopathy. Key words: cervical myelopathy outcome, cross-cultural adaptation, grading scale, JOACMEQ, mJOA.
Level of Evidence: N/A Spine 2014;39:963-970
C
ervical spondylotic myelopathy (CSM) is the tnost common cause of spinal cord dysfunction in elderly adults worldwide'-^ and is becoming one of the most common disease treated by spine surgeons. ' Civen its increasing prevalence and significant impact on quality of life (QOL), the ability to compare and combine various evaluation and outcome results globally is a necessity for successful discussion on the management of CSM. A number of grading scales have been developed in the last 2 decades to assist clinicians to conduct a more systematic outcome assessment for CSM. Of all the questionnaires, the modified Japanese Orthopaedic Association (mJOA) score is one of the most commonly used and established in the literature,'* with better sensitivity and specificity than other tools www.spinejournal.com
963
Spine
HEALTH SERVICES RESEARCH
reported in the literature.^-* In a recently published consensus statement, the authors strongly recommended that the Neck Disability Index (NDI) and the mJOA score be used routinely on all patients with CSM as an evaluation and outcome measure.'' Despite the wide acceptance of the mJOA worldwide, the authors of the original JOA further developed the mJOA in 1999 with an emphasis on assessing patients' satisfaction, disability, handicaps, and general health from their own perspective. The updated mJOA is called the "Japanese Orthopaedic Association Cervical Myelopathy Evaluation Questionnaire (JOACMEQ)."" JOACMEQ has been demonstrated to have good validity as well as reliability and correlates strongly with other commonly used neck pain and QOL questionnaires.'" Despite the well-established evidence and growing support advocating the use of the mJOA and subsequently the JOACMEQ internationally, at the time of this writing, there is no Chinese version of the mJOA or the JOACMEQ available for use in the Chinese-speaking population. Given the equally high prevalence rate of CSM in Asia''' and the importance of having questionnaires in the native language of the patients,'^'"^ there is an urgent need to develop a Chinese version of the JOACMEQ. Development of such tool will not only allow surgeons and health professionals treating CSM to gain access to a globally accepted outcome measure but also, more importantly, allows better standardizing of clinical practice and international research. In this article, the authors report on the development of a Chinese version of the JOACMEQ following the internationally accepted guideline for the cultural adaptation of a self-reporting questionnaire by Beaton et al.^^ Apart from a detailed description of the translation process, results of the statistical tests assessing the psychometric properties of the translated Chinese version are also presented. The purpose of this study is, therefore, to generate a Chinese version of the JOACMEQ that is semantically and conceptually as close as possible to the original JOACMEQ, which is appropriate for the Chinese-speaking population.
MATERIALS AND METHODS Design The study was conducted in 2 stages: the first stage included translation and cross-cultural adaptation of the English JOACMEQ including a small pilot study to synthesize a provisional Chinese JOACMEQ. The second stage involved the testing of the psychometric properties of the final Chinese JOACMEQ in a cross-sectional validation study.
Participants This study involved 2 university teaching hospitals located in 2 different geographic areas representing slightly different social and cultural contexts of the greater region. Patients who attended the department of surgery's outpatient clinics from August 2013 to December 2013 with a predominant complain of neck pain were recruited for the study. Given the primary goal of this study was the validation of the translation 964
www.spinejournal.conr)
Cross-cultural Adaptation of JOACMEQ • Chien et al
of the JOACMEQ, the study sample included but not limited to patients with a clinically confirmed CSM. All patients were assessed by the same neurosurgeon, and the diagnosis of CSM was confirmed clinically if the sagittal plane T2WI of cervical spine magnetic resonance image showed absence of cerebrospinal fluid signal with or without cord edema and/or atrophy on the level of pathology with buateral hand numbness and/ or leg weakness with increased deep tendon reflexes.'* Baseline demographics and patient characteristics were collected as part of the hospital admission process. The participants were asked to complete the questionnaires (Chinese JOACMEQ and NDI) on 2 separate occasions. The initial questionnaires were completed at the time of attending their consultation with their respective physicians, and the second set of questionnaires were provided in a reply-paid envelope, and patients were asked to complete these at home and return them a week later. All participants gave their written consent, and the institution's ethical committee granted the ethical approval for the study.
The Translation and Adaptation Process Forward Translation
Two native Chinese speakers with fluency in English first performed forward translation of the original JOACMEQ into Chinese. The 2 Chinese versions produced by the each translator were reviewed and compared by a third independent assessor who is also fluent in both languages. A pretesting Chinese version of the JOACMEQ was then synthesized after a consensus was reached. Backward Translation
Working from the pretesting Chinese JOACMEQ and remaining totally blinded to the original version, 2 bilingual translators whose native language was English translated the pretesting Chinese version back into English. Group Meeting All of the authors and the translators then meet to consider all of the versions of the questionnaires. Each question was discussed and updated until the general consensus was reached, with a special emphasis on determining the most appropriate wording for the targeted elderly population. The provisional Chinese version was then synthesized and used for pretesting.
Participant Pretest To determine whether the translation (instructions, items, and response choices) is acceptable for the general clinical population, the provisional version was firstly administered to 10 participants to investigate whether it is understood by the way it is supposed to be and whether the language used is appropriate without confusions. After the completion of the initial questionnaires, face-toface interviews were also conducted to ask about whether the participant had any difficulty in understanding the questionnaire and check the patient's interpretation of all items. Although most of the participants were able to complete the May 2014
J Spine
HEALTH SERVICES RESEARCH
questionnaire without any difficulties, 3 patients indicated some difficulty with Question 2-2. The particular question asked the patients whether they are able to eat a meal with their dominant hand using a spoon or a fork, but the patients indicated that they rarely have the need to use fork for eating meals. Given that chopsticks are the most popular eating utensils of choice in the targeted Chinese-speaking population, consensus was reached among the authors to replace the wording of "a spoon or a fork" with "chopsticks or a spoon" for Question 2-2. The provisional Chinese JOACMEQ was then updated and the final Chinese JOACMEQ version (refer to Supplemental Digital Content Appendix, available at http:// links.lww.com/BRS/A869) synthesized accordingly for testing.
Cross-cultural Adaptation of JOACMEQ • Chien et al
Discriminant Validity To determine whether the Chinese JOACMEQ would be able to accurately discriminate between patients with clinically diagnosed CSM and those with neck pain of other causes, the optimal cutoff point was calculated using the receiver operating characteristic (ROC) curve on the basis of the Youden Index to determine the associated sensitivity and specificity.
Statistical Analysis Statistical analysis was performed using PASW Statistics 22 for Macintosh (SPSS, Chicago, IL). Where applicable, a P value of less than 0.05 was considered significant. RESULTS
Psychometric Testing Construct Validity (Definition of Domains) An exploratory factor analysis (principal component analysis with varimax rotation) was carried out to investigate the correlations among the Chinese JOACMEQ questions and to compare the factorial structure of the output with those defined by the original JOACMEQ. A component is considered significant if its eigenvalue is above 1.0. Given the original JOACMEQ defined the 24 questions into 5 domains for scoring, a second factor analysis with 5 factors predesignated was also carried out. Internal Consistency Internal consistency of the Chinese JOACMEQ was evaluated using Cronbach a coefficient. Based on the generally accepted standard in the literature, a greater than 0.9 was regarded as excellent; ct of 0.8 or greater was regarded as good, and a of 0.7 or greater was regarded as acceptable and less than 0.5 as weak." Reproducibility (Test-Retest Reliability) The Chinese JOACMEQ was administered to the same group of patients on 2 different occasions approximately a week apart. This standard time frame was chosen because it is unlikely that the disease status would have changed in this short period but the time frame is long enough to avoid recall of responses. Intraclass correlation coefficient (ICC) together with the 95% confidence intervals was used to evaluate the agreement between the test and the retest JOACMEQ individual domain scores. Based on the commonly accepted literature, ICC of greater than 0.75 is considered as good reliability and ICC of greater than 0.85 as excellent reliability.-" Concurrent Validity The concurrent validity (clinical validity) was assessed by calculating the Spearman correlation coefficient (Spearman p) to determine the correlation between the JOACMEQ individual domain scores against the Chinese NDI. In general, p = 0.1 to 0.2 was regarded as a weak or no association, p = 0.2 to 0.4 as a weak association, p = 0.4 to 0.6 as a moderate relationship, and p = 0.6 to 0.8 as a strong association. Spine
Participants A total of 105 patients (63 males and 42 females) with the mean age of 51.5 ± 15.8 years (range, 21-79 yr) were included in the study. Ofthe 105 patients, 6 (5.7%) received no formal education and 22 (21.0%), 46 (43.8%), and 31 (29.5%) completed junior high school, high school, and university education, respectively. Ninety-eight of the patients (response rate = 93%) completed all of the questionnaires on both occasions. Of the 98 patients, 26 (27%) were clinically diagnosed with CSM. Independent samples t test revealed a significantly lower upper and lower extremity domain scores in those diagnosed with CSM (upper extremity: 82.1 ± 18.5 vs. 92.1 ± 11.0, P = 0.028; lower extremity: 74 ± 31.2 vs. 88.0 ± 15.7, P = 0.024). No significant difference was found for the other Chinese JOACMEQ domain scores and the NDI between the groups. A summary of the JOACMEQ and NDI scores is presented in Figure 1.
Myelopathy vs. Non myelopathy Patients
Cervical spine function
Upper extremity Lower extremity Bladder function function function
«Myelopathy is Nonmyelopathy
Quaiity of Life
Chinese NDI and JOACMEQ Domains
Figure 1. A comparison between patients with clinically diagnosed cervical myelopathy and those with other causes of neck pain using the Chinese NDI and Chinese version of JOACMEQ domain scores. NDI indicates Neck Disability Index; JOACMEQ, Japanese Orthopaedic Association Cervical Myelopathy Evaluation Questionnaire. www.spinejournal.com
965
Spine
Cross-cultural Adaptation of JOACMEQ • Chien et al
HEALTH SERVICES RESEARCH
The mean time to complete the questionnaire was 3 minutes (range: 1-6 min) for the NDI and 8 minutes (range: 4-15 min) for the JOACMFQ. Twelve patients required help reading some or all of the questions and 3 patients required assistance in writing the responses. Four of the returned JOACMFQ questionnaires were found to contain missing responses for at least 1 of the questions.
Function, and QOL Domains in the original JOACMFQ received the greatest absolute factor loading within their respective domain. However, Question 5-8 was found to have significant contribution to not only the QOL domain that it was designed to represent but also the Upper Extremity and Bladder Function Domains. The contribution of the individual questions to each of the component is presented in Table 2.
Psychometric Properties Construct Validity On factor analysis, 24 questions in the Chinese JOACMEQ were allocated into 6 components with eigenvalues above 1 and jointly accounted for 6 3 % of the total variance. The results are shown in Table 1. When the 24 questions were forced into 5 factors, it was noted that all of the questions allocated to the Cervical Spine Function, Lower Fxtremity
Contribution Cumulative Contribution Rate, % Factor Eigenvalue Rate, % 1
7.05
29.4
29.4
2
2.25
9.4
38.8
3
1.94
8.1
46.9
4
1.37
5.7
52.6
5
1.28
5.4
57.9
6
1.11
4.6
62.5
7
0.97
4.0
66.6
8
0.93
3.9
70.4
9
0.85
3.5
74.0
10
0.68
2.8
76.8
11
0.62
2.6
79.4
12
0.60
2.5
81.9
13
0.56
2.3
84.2
14
0.50
2.1
86.3
15
0.50
2.1
88.4
16
0.44
1.8
90.2
17
0.39
1.6
91.8
18
0.37
1.5
93.3
19
0.34
1.4
94.8
20
0.32
1.3
96.1
21
0.27
1.1
97.2
22
0.25
1.0
98.3
23
0.22
0.9
99.2
24
0.20
0.8
100.0
*n = 173. Values in boldface indicate eigenvalues greater tiian 1.0
966
www.spinejournal.com
Internal Consistency The Cronbach a obtained for the total and each of the 5 domains is presented in Table 3. The overall internal consistency of the Chinese JOACMEQ with the 24 items combined was found to be 0.88, indicating good to excellent internal consistency of the questionnaire. For the domain scores, the a values were weak to acceptable for upper extremity and bladder function (a = 0.65 and 0.56, respectively), acceptable for cervical spine function (ce = 0.72), and good for lower extremity and QOL (a = 0.84 and 0.83). Reproducibility For the domain scores, excellent ICC scores were observed for the lower extremity domain (ICC = 0.90, P = 0.6), followed by the QOL, Bladder, and Upper Extremity Domains (ICC = 0.87, P = 0.323; ICC = 0.87, P = 0.168; and ICC = 0.85, P = 0.94, respectively). A good ICC score was observed for the Cervical Spine Function (ICC = 0.79, P = 0.35). Chinese JOACMEQ test-retest subdomain scores are presented in Figure 2, and a summary of the ICC scores is presented in Table 3. Concurrent Validity Clinical validity was assessed by the Spearman correlation coefficient of the Chinese JOACMEQ individual domain scores against the Chinese NDI score. As shown in Table 3, a significant negative weak to moderate correlation was found for all of the domain scores (Spearman p range = - 0 . 4 7 to -0.54, P < 0.01), with the exception of the QOL domain where a significant strong association (Spearman p = —0.76, P < 0.01) was identified. Discriminant Validity The area under the ROC curve for the individual domain scores ranged from 0.46 to 0.82, indicating weak to good discriminating power. The upper and lower extremity function domains reached statistical significance, with the area under the ROC curve found to be 0.75 (P = 0.04) and 0.82 (P = 0.02), respectively. With an optimal cutoff point of 72 for the upper extremity function, the sensitivity was 96% and specificity was 8 1 % . A cutoff point of 62 for the lower extremity score achieved a sensitivity of 94% and specificity of 73%. Figure 3 shows the ROC curve for the Upper and Lower Fxtremity Domain scores.
DISCUSSION This study aimed to translate the English JOACMEQ into Chinese while maintaining the psychometric properties of the May 2014
Spine
HEALTH SERVICES RESEARCH
Cross-cultural Adaptation of JOACMEQ • Chien et al
|ms2|2^pfimax-Kotated Factor Matrix of the Ch nese version ot the Japanese Urthopaedic J ^ H B ä^iT-' ' " Question
1. Cervical Spine
"
2. Upper Limb
3. Lower Limb
4. Bladder
5. QOL
0.832*
0.098
-0.004
0.030
-0.085
Ql-2
0.590*
"0.124
0.075
-0.060
0.067
Ql-3
0.702*
0.284
0.061
0.164
0.215
Ql-4
0.614*
0.312*
0.251
0.095
0.164
Q2-1
0.140
0.772*
0.252
0.113
0.001
Q2-2
0.059
0.827*
0.200
-0.017
0.123
Q2-3
0.287
0.503*
0.163
0.066
0.116
Q3-1
0.048
0.413*
0.683*
0.216
0.053
Q3-2
0.017
0.204
0.728*
-0.129
0.261
Q3-3
0.061
0.081
0.838*
0.125
0.048
Q3-4
0.181
0.107
0.671*
0.176
0.216
Q3-5
0.072
0.156
0.794*
0.102
0.201
Q4-1
0.108
-0.118
0.421
0.586*
0.002
Q4-2
0.252
-0.178
0.334
0.593*
-0.136
Q4-3
-0.007
-0.016
0.071
0.693*
0.312
Q4-4
0.042
0.110
0.527
0.460*
0.046
Q5-1
0.143
0.021
0.202
0.372
0.446*
Q5-2
0.218
0.170
0.487
-0.022
0.531*
Q5-3
0.369
0.134
0.203
-0.162
0.669*
Q5-4
0.069
-0.045
0.160
0.073
0.734*
Q5-5
0.222
0.074
0.013
0.066
0.840*
Q5-6
0.019
0.129
0.182
0.124
0.726*
Q5-7
0.065
0.031
0.185
0.291
0.543*
Q5-8
-0.102
0.564*
0.495*
Ql-1
0.353*
-0.018
The questions found to have the highest loading for each of the subdomains. QOL indicates quality of life.
original JOACMFQ. The results indicated that the Chinese JOACMEQ is comparable with the original JOACMEQ and could be used in the future to detect disability and functional limitations because of CSM in Chinese populations. Construct validity is a major criterion of validity for any self-reporting questionnaire. In the current study, the construct validity of the Chinese JOACMEQ was assessed utilizing factorial analysis. The initial factor analysis revealed that the cumulative contribution rates from the top 6 factors identified in the Chinese JOACMEQ were comparable with those published in the original JOACMFQ.'- This result is indicative of the similar factorial structure and construct validity between the 2 questionnaires. Utilizing the same factor analysis protocol carried out by the original JOACMEQ'- in which the 24 questions were forced into 5 factors, all questions, with the exception of Question 5-8, were allocated correctly into Spine
their respective domains (based on the allocation and scoring system of the original JOACMEQ). Question 5-8, which contributed significantly to the QOL domain that it was designed to represent, was also found to have significant contribution to the upper extremity and bladder function domains. The authors speculate that the nature of the question itself may explain its contribution to the multiple domains. More specifically. Question 5-8 asked the patient "Do you feel your health will get worse?". Such question may be considered as a generalized question attempting to gauge patients' subjective expectation of their own health recovery, and in order to respond to such question, patients will inevitably consider and evaluate their current symptoms and their severity. It is, therefore, hypothesized that patients reporting higher levels of upper extremity and bladder dysfunction may be more likely to also report worse expectation of their health recovery and www.spinejournal.com
967
Spine
Cross-cultural Adaptation of JOACMEQ • Chien et al
HEALTH SERVICES RESEARCH
J ^ ^ ^ ^ l ^ r o n b a c h a Coefficient, Intraclass Correlation Coefficient, and Spearman p values ofJJ Cronbach a Coefficient
Intraclass Correlation Coefficient (P)
Spearman p (P)
JOACMEQ—24 Questions
0.884
Cervical spine function
0.723
0.793 (0.35)
-0.465 (
HEALTH SERVICES RESEARCH
Translation, Cross-cultural Adaptation, and Validation of a Chinese Version of the Japanese Orthopaedic Association Cervical Myelopathy Evaluation Questionnaire Andy Chien, PhD,*+ Dar-Ming Lai, PhD,* Chih-Hsiu Cheng, PhD,§ Shwu-Fen Wang, PhD,t Wei-Li Hsu, PhD,t and Jaw-Lin Wang, PhD*
Study Design. Translation and psychometric testing of a questionnaire. Objective. To adapt the Japanese Orthopaedic Association Cervical Myelopathy Evaluation Questionnaire QOACMEQ) crossculturally to Chinese language and to investigate the psychometric properties ofthe translated Chinese lOACMEQ. Summary of Bacitground Data. The recently developed JOACMEQ is a self-reported questionnaire used to assess the severity of the cervical spondylotic myelopathy. Despite its acceptance in the literature, the JOACMEQ has not been translated and validated for Chinese-speaking patients. M e t h o d s . Translation and cross-cultural adaptation of the original JOACMEQ was performed according to published guidelines. A total of 105 patients with a predominant complain of neck pain that included but not limited to patients with a clinically confirmed cervical spondylotic myelopathy were included. Psychometric testing of the responses included construct validity (factor analysis), internal consistency (Cronbach a), test-retest reproducibüity (intraclass correlation coefficient), and concurrent validity (Spearman p). The discriminatory power of differentiating those with myelopathy from
From the 'Institute of Biomédical Engineering, National Taiwan University, Taipei, Taiwan; tSchool and Graduate Institute of Physical Therapy, College of Medicine, National Taiwan University, Taipei, Taiwan; ^Department of Surgery, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan; and §Department of Physical Therapy and Graduate Institute of Rehabilitation Science, Chang Gung University, Taoyuan, Taiwan. Acknowledgment date: January 2, 2014. Revision date: February 19, 2014. Acceptance date: February 24, 2014. The manuscript submitted does not contain information about medical device(s)/drug(s). The National Science Council, Taiwan (NSC 101-2221-F-002-060-MY3) grant funds were received to support this work. No relevant financial activities outside the submitted work. Address correspondence and reprint requests to |aw-Lin Wang, PhD, Department of Mechanical Engineering, College of Engineering, National Taiwan University, 602 |en-Su Hall, 1 Section 4, Roosevelt Rd, Taipei 10617, Taiwan, ROC; E-mail: [email protected] DOI: 10.1097/BRS.0000000000000313 Spine
those without was assessed by the area under the receiver operating characteristic curve and with the optimal cutoff points determined. Results. Construct validity testing revealed a comparable factorial structure of the Chinese JOACMEQ with the original JOACMEQ. The Cronbach a of the Chinese JOACMEQ was 0.88, indicating good internal consistency. Good to excellent test-retest reliability was demonstrated for all 5 domain scores (intraclass correlation coefficient range = 0.79-0.90). A significant strong association between the Chinese Neck Disability Index and the quality of life domain of the Chinese JOACMEQ (Spearman p = - 0 . 7 6 , P < 0.01 ) was identified. The upper and lower extremity function domain scores were found to be significant, with the area under the receiver operating characteristic curve found to be 0.75 {P = 0.04) and 0.82 (P = 0.02), respectively. Conclusion. The results indicated that the Chinese JOACMEQ successfully retained the psychometric properties of the original JOACMEQ and support the usefulness of the Chinese JOACMEQ as an appropriate supplementary diagnostic and outcome measure for Chinese patients suspected of cervical spondylotic myelopathy. Key words: cervical myelopathy outcome, cross-cultural adaptation, grading scale, JOACMEQ, mJOA.
Level of Evidence: N/A Spine 2014;39:963-970
C
ervical spondylotic myelopathy (CSM) is the tnost common cause of spinal cord dysfunction in elderly adults worldwide'-^ and is becoming one of the most common disease treated by spine surgeons. ' Civen its increasing prevalence and significant impact on quality of life (QOL), the ability to compare and combine various evaluation and outcome results globally is a necessity for successful discussion on the management of CSM. A number of grading scales have been developed in the last 2 decades to assist clinicians to conduct a more systematic outcome assessment for CSM. Of all the questionnaires, the modified Japanese Orthopaedic Association (mJOA) score is one of the most commonly used and established in the literature,'* with better sensitivity and specificity than other tools www.spinejournal.com
963
Spine
HEALTH SERVICES RESEARCH
reported in the literature.^-* In a recently published consensus statement, the authors strongly recommended that the Neck Disability Index (NDI) and the mJOA score be used routinely on all patients with CSM as an evaluation and outcome measure.'' Despite the wide acceptance of the mJOA worldwide, the authors of the original JOA further developed the mJOA in 1999 with an emphasis on assessing patients' satisfaction, disability, handicaps, and general health from their own perspective. The updated mJOA is called the "Japanese Orthopaedic Association Cervical Myelopathy Evaluation Questionnaire (JOACMEQ)."" JOACMEQ has been demonstrated to have good validity as well as reliability and correlates strongly with other commonly used neck pain and QOL questionnaires.'" Despite the well-established evidence and growing support advocating the use of the mJOA and subsequently the JOACMEQ internationally, at the time of this writing, there is no Chinese version of the mJOA or the JOACMEQ available for use in the Chinese-speaking population. Given the equally high prevalence rate of CSM in Asia''' and the importance of having questionnaires in the native language of the patients,'^'"^ there is an urgent need to develop a Chinese version of the JOACMEQ. Development of such tool will not only allow surgeons and health professionals treating CSM to gain access to a globally accepted outcome measure but also, more importantly, allows better standardizing of clinical practice and international research. In this article, the authors report on the development of a Chinese version of the JOACMEQ following the internationally accepted guideline for the cultural adaptation of a self-reporting questionnaire by Beaton et al.^^ Apart from a detailed description of the translation process, results of the statistical tests assessing the psychometric properties of the translated Chinese version are also presented. The purpose of this study is, therefore, to generate a Chinese version of the JOACMEQ that is semantically and conceptually as close as possible to the original JOACMEQ, which is appropriate for the Chinese-speaking population.
MATERIALS AND METHODS Design The study was conducted in 2 stages: the first stage included translation and cross-cultural adaptation of the English JOACMEQ including a small pilot study to synthesize a provisional Chinese JOACMEQ. The second stage involved the testing of the psychometric properties of the final Chinese JOACMEQ in a cross-sectional validation study.
Participants This study involved 2 university teaching hospitals located in 2 different geographic areas representing slightly different social and cultural contexts of the greater region. Patients who attended the department of surgery's outpatient clinics from August 2013 to December 2013 with a predominant complain of neck pain were recruited for the study. Given the primary goal of this study was the validation of the translation 964
www.spinejournal.conr)
Cross-cultural Adaptation of JOACMEQ • Chien et al
of the JOACMEQ, the study sample included but not limited to patients with a clinically confirmed CSM. All patients were assessed by the same neurosurgeon, and the diagnosis of CSM was confirmed clinically if the sagittal plane T2WI of cervical spine magnetic resonance image showed absence of cerebrospinal fluid signal with or without cord edema and/or atrophy on the level of pathology with buateral hand numbness and/ or leg weakness with increased deep tendon reflexes.'* Baseline demographics and patient characteristics were collected as part of the hospital admission process. The participants were asked to complete the questionnaires (Chinese JOACMEQ and NDI) on 2 separate occasions. The initial questionnaires were completed at the time of attending their consultation with their respective physicians, and the second set of questionnaires were provided in a reply-paid envelope, and patients were asked to complete these at home and return them a week later. All participants gave their written consent, and the institution's ethical committee granted the ethical approval for the study.
The Translation and Adaptation Process Forward Translation
Two native Chinese speakers with fluency in English first performed forward translation of the original JOACMEQ into Chinese. The 2 Chinese versions produced by the each translator were reviewed and compared by a third independent assessor who is also fluent in both languages. A pretesting Chinese version of the JOACMEQ was then synthesized after a consensus was reached. Backward Translation
Working from the pretesting Chinese JOACMEQ and remaining totally blinded to the original version, 2 bilingual translators whose native language was English translated the pretesting Chinese version back into English. Group Meeting All of the authors and the translators then meet to consider all of the versions of the questionnaires. Each question was discussed and updated until the general consensus was reached, with a special emphasis on determining the most appropriate wording for the targeted elderly population. The provisional Chinese version was then synthesized and used for pretesting.
Participant Pretest To determine whether the translation (instructions, items, and response choices) is acceptable for the general clinical population, the provisional version was firstly administered to 10 participants to investigate whether it is understood by the way it is supposed to be and whether the language used is appropriate without confusions. After the completion of the initial questionnaires, face-toface interviews were also conducted to ask about whether the participant had any difficulty in understanding the questionnaire and check the patient's interpretation of all items. Although most of the participants were able to complete the May 2014
J Spine
HEALTH SERVICES RESEARCH
questionnaire without any difficulties, 3 patients indicated some difficulty with Question 2-2. The particular question asked the patients whether they are able to eat a meal with their dominant hand using a spoon or a fork, but the patients indicated that they rarely have the need to use fork for eating meals. Given that chopsticks are the most popular eating utensils of choice in the targeted Chinese-speaking population, consensus was reached among the authors to replace the wording of "a spoon or a fork" with "chopsticks or a spoon" for Question 2-2. The provisional Chinese JOACMEQ was then updated and the final Chinese JOACMEQ version (refer to Supplemental Digital Content Appendix, available at http:// links.lww.com/BRS/A869) synthesized accordingly for testing.
Cross-cultural Adaptation of JOACMEQ • Chien et al
Discriminant Validity To determine whether the Chinese JOACMEQ would be able to accurately discriminate between patients with clinically diagnosed CSM and those with neck pain of other causes, the optimal cutoff point was calculated using the receiver operating characteristic (ROC) curve on the basis of the Youden Index to determine the associated sensitivity and specificity.
Statistical Analysis Statistical analysis was performed using PASW Statistics 22 for Macintosh (SPSS, Chicago, IL). Where applicable, a P value of less than 0.05 was considered significant. RESULTS
Psychometric Testing Construct Validity (Definition of Domains) An exploratory factor analysis (principal component analysis with varimax rotation) was carried out to investigate the correlations among the Chinese JOACMEQ questions and to compare the factorial structure of the output with those defined by the original JOACMEQ. A component is considered significant if its eigenvalue is above 1.0. Given the original JOACMEQ defined the 24 questions into 5 domains for scoring, a second factor analysis with 5 factors predesignated was also carried out. Internal Consistency Internal consistency of the Chinese JOACMEQ was evaluated using Cronbach a coefficient. Based on the generally accepted standard in the literature, a greater than 0.9 was regarded as excellent; ct of 0.8 or greater was regarded as good, and a of 0.7 or greater was regarded as acceptable and less than 0.5 as weak." Reproducibility (Test-Retest Reliability) The Chinese JOACMEQ was administered to the same group of patients on 2 different occasions approximately a week apart. This standard time frame was chosen because it is unlikely that the disease status would have changed in this short period but the time frame is long enough to avoid recall of responses. Intraclass correlation coefficient (ICC) together with the 95% confidence intervals was used to evaluate the agreement between the test and the retest JOACMEQ individual domain scores. Based on the commonly accepted literature, ICC of greater than 0.75 is considered as good reliability and ICC of greater than 0.85 as excellent reliability.-" Concurrent Validity The concurrent validity (clinical validity) was assessed by calculating the Spearman correlation coefficient (Spearman p) to determine the correlation between the JOACMEQ individual domain scores against the Chinese NDI. In general, p = 0.1 to 0.2 was regarded as a weak or no association, p = 0.2 to 0.4 as a weak association, p = 0.4 to 0.6 as a moderate relationship, and p = 0.6 to 0.8 as a strong association. Spine
Participants A total of 105 patients (63 males and 42 females) with the mean age of 51.5 ± 15.8 years (range, 21-79 yr) were included in the study. Ofthe 105 patients, 6 (5.7%) received no formal education and 22 (21.0%), 46 (43.8%), and 31 (29.5%) completed junior high school, high school, and university education, respectively. Ninety-eight of the patients (response rate = 93%) completed all of the questionnaires on both occasions. Of the 98 patients, 26 (27%) were clinically diagnosed with CSM. Independent samples t test revealed a significantly lower upper and lower extremity domain scores in those diagnosed with CSM (upper extremity: 82.1 ± 18.5 vs. 92.1 ± 11.0, P = 0.028; lower extremity: 74 ± 31.2 vs. 88.0 ± 15.7, P = 0.024). No significant difference was found for the other Chinese JOACMEQ domain scores and the NDI between the groups. A summary of the JOACMEQ and NDI scores is presented in Figure 1.
Myelopathy vs. Non myelopathy Patients
Cervical spine function
Upper extremity Lower extremity Bladder function function function
«Myelopathy is Nonmyelopathy
Quaiity of Life
Chinese NDI and JOACMEQ Domains
Figure 1. A comparison between patients with clinically diagnosed cervical myelopathy and those with other causes of neck pain using the Chinese NDI and Chinese version of JOACMEQ domain scores. NDI indicates Neck Disability Index; JOACMEQ, Japanese Orthopaedic Association Cervical Myelopathy Evaluation Questionnaire. www.spinejournal.com
965
Spine
Cross-cultural Adaptation of JOACMEQ • Chien et al
HEALTH SERVICES RESEARCH
The mean time to complete the questionnaire was 3 minutes (range: 1-6 min) for the NDI and 8 minutes (range: 4-15 min) for the JOACMFQ. Twelve patients required help reading some or all of the questions and 3 patients required assistance in writing the responses. Four of the returned JOACMFQ questionnaires were found to contain missing responses for at least 1 of the questions.
Function, and QOL Domains in the original JOACMFQ received the greatest absolute factor loading within their respective domain. However, Question 5-8 was found to have significant contribution to not only the QOL domain that it was designed to represent but also the Upper Extremity and Bladder Function Domains. The contribution of the individual questions to each of the component is presented in Table 2.
Psychometric Properties Construct Validity On factor analysis, 24 questions in the Chinese JOACMEQ were allocated into 6 components with eigenvalues above 1 and jointly accounted for 6 3 % of the total variance. The results are shown in Table 1. When the 24 questions were forced into 5 factors, it was noted that all of the questions allocated to the Cervical Spine Function, Lower Fxtremity
Contribution Cumulative Contribution Rate, % Factor Eigenvalue Rate, % 1
7.05
29.4
29.4
2
2.25
9.4
38.8
3
1.94
8.1
46.9
4
1.37
5.7
52.6
5
1.28
5.4
57.9
6
1.11
4.6
62.5
7
0.97
4.0
66.6
8
0.93
3.9
70.4
9
0.85
3.5
74.0
10
0.68
2.8
76.8
11
0.62
2.6
79.4
12
0.60
2.5
81.9
13
0.56
2.3
84.2
14
0.50
2.1
86.3
15
0.50
2.1
88.4
16
0.44
1.8
90.2
17
0.39
1.6
91.8
18
0.37
1.5
93.3
19
0.34
1.4
94.8
20
0.32
1.3
96.1
21
0.27
1.1
97.2
22
0.25
1.0
98.3
23
0.22
0.9
99.2
24
0.20
0.8
100.0
*n = 173. Values in boldface indicate eigenvalues greater tiian 1.0
966
www.spinejournal.com
Internal Consistency The Cronbach a obtained for the total and each of the 5 domains is presented in Table 3. The overall internal consistency of the Chinese JOACMEQ with the 24 items combined was found to be 0.88, indicating good to excellent internal consistency of the questionnaire. For the domain scores, the a values were weak to acceptable for upper extremity and bladder function (a = 0.65 and 0.56, respectively), acceptable for cervical spine function (ce = 0.72), and good for lower extremity and QOL (a = 0.84 and 0.83). Reproducibility For the domain scores, excellent ICC scores were observed for the lower extremity domain (ICC = 0.90, P = 0.6), followed by the QOL, Bladder, and Upper Extremity Domains (ICC = 0.87, P = 0.323; ICC = 0.87, P = 0.168; and ICC = 0.85, P = 0.94, respectively). A good ICC score was observed for the Cervical Spine Function (ICC = 0.79, P = 0.35). Chinese JOACMEQ test-retest subdomain scores are presented in Figure 2, and a summary of the ICC scores is presented in Table 3. Concurrent Validity Clinical validity was assessed by the Spearman correlation coefficient of the Chinese JOACMEQ individual domain scores against the Chinese NDI score. As shown in Table 3, a significant negative weak to moderate correlation was found for all of the domain scores (Spearman p range = - 0 . 4 7 to -0.54, P < 0.01), with the exception of the QOL domain where a significant strong association (Spearman p = —0.76, P < 0.01) was identified. Discriminant Validity The area under the ROC curve for the individual domain scores ranged from 0.46 to 0.82, indicating weak to good discriminating power. The upper and lower extremity function domains reached statistical significance, with the area under the ROC curve found to be 0.75 (P = 0.04) and 0.82 (P = 0.02), respectively. With an optimal cutoff point of 72 for the upper extremity function, the sensitivity was 96% and specificity was 8 1 % . A cutoff point of 62 for the lower extremity score achieved a sensitivity of 94% and specificity of 73%. Figure 3 shows the ROC curve for the Upper and Lower Fxtremity Domain scores.
DISCUSSION This study aimed to translate the English JOACMEQ into Chinese while maintaining the psychometric properties of the May 2014
Spine
HEALTH SERVICES RESEARCH
Cross-cultural Adaptation of JOACMEQ • Chien et al
|ms2|2^pfimax-Kotated Factor Matrix of the Ch nese version ot the Japanese Urthopaedic J ^ H B ä^iT-' ' " Question
1. Cervical Spine
"
2. Upper Limb
3. Lower Limb
4. Bladder
5. QOL
0.832*
0.098
-0.004
0.030
-0.085
Ql-2
0.590*
"0.124
0.075
-0.060
0.067
Ql-3
0.702*
0.284
0.061
0.164
0.215
Ql-4
0.614*
0.312*
0.251
0.095
0.164
Q2-1
0.140
0.772*
0.252
0.113
0.001
Q2-2
0.059
0.827*
0.200
-0.017
0.123
Q2-3
0.287
0.503*
0.163
0.066
0.116
Q3-1
0.048
0.413*
0.683*
0.216
0.053
Q3-2
0.017
0.204
0.728*
-0.129
0.261
Q3-3
0.061
0.081
0.838*
0.125
0.048
Q3-4
0.181
0.107
0.671*
0.176
0.216
Q3-5
0.072
0.156
0.794*
0.102
0.201
Q4-1
0.108
-0.118
0.421
0.586*
0.002
Q4-2
0.252
-0.178
0.334
0.593*
-0.136
Q4-3
-0.007
-0.016
0.071
0.693*
0.312
Q4-4
0.042
0.110
0.527
0.460*
0.046
Q5-1
0.143
0.021
0.202
0.372
0.446*
Q5-2
0.218
0.170
0.487
-0.022
0.531*
Q5-3
0.369
0.134
0.203
-0.162
0.669*
Q5-4
0.069
-0.045
0.160
0.073
0.734*
Q5-5
0.222
0.074
0.013
0.066
0.840*
Q5-6
0.019
0.129
0.182
0.124
0.726*
Q5-7
0.065
0.031
0.185
0.291
0.543*
Q5-8
-0.102
0.564*
0.495*
Ql-1
0.353*
-0.018
The questions found to have the highest loading for each of the subdomains. QOL indicates quality of life.
original JOACMFQ. The results indicated that the Chinese JOACMEQ is comparable with the original JOACMEQ and could be used in the future to detect disability and functional limitations because of CSM in Chinese populations. Construct validity is a major criterion of validity for any self-reporting questionnaire. In the current study, the construct validity of the Chinese JOACMEQ was assessed utilizing factorial analysis. The initial factor analysis revealed that the cumulative contribution rates from the top 6 factors identified in the Chinese JOACMEQ were comparable with those published in the original JOACMFQ.'- This result is indicative of the similar factorial structure and construct validity between the 2 questionnaires. Utilizing the same factor analysis protocol carried out by the original JOACMEQ'- in which the 24 questions were forced into 5 factors, all questions, with the exception of Question 5-8, were allocated correctly into Spine
their respective domains (based on the allocation and scoring system of the original JOACMEQ). Question 5-8, which contributed significantly to the QOL domain that it was designed to represent, was also found to have significant contribution to the upper extremity and bladder function domains. The authors speculate that the nature of the question itself may explain its contribution to the multiple domains. More specifically. Question 5-8 asked the patient "Do you feel your health will get worse?". Such question may be considered as a generalized question attempting to gauge patients' subjective expectation of their own health recovery, and in order to respond to such question, patients will inevitably consider and evaluate their current symptoms and their severity. It is, therefore, hypothesized that patients reporting higher levels of upper extremity and bladder dysfunction may be more likely to also report worse expectation of their health recovery and www.spinejournal.com
967
Spine
Cross-cultural Adaptation of JOACMEQ • Chien et al
HEALTH SERVICES RESEARCH
J ^ ^ ^ ^ l ^ r o n b a c h a Coefficient, Intraclass Correlation Coefficient, and Spearman p values ofJJ Cronbach a Coefficient
Intraclass Correlation Coefficient (P)
Spearman p (P)
JOACMEQ—24 Questions
0.884
Cervical spine function
0.723
0.793 (0.35)
-0.465 (
